X-ray sources have been in use for medical and physics research applications since the discovery of x-rays by William Roentgen at the turn of the century. Early sources were based on acceleration of electrons in an x-ray vacuum tube (using high voltage power supplies) and their subsequent collision with a cooled metal anode, which produced K- and L-edge x-ray line emission from the metal (usually copper) atoms in addition to continuum radiation from Bremsstrahlung radiation. The shape and duration of the x-ray pulse was limited by the inductive rise time of the electron tube circuitry (in the microsecond range) for the pulse leading edge, and the power supply and cooling capability for the pulse duration.
With the advent of the electron beam accelerators in the 1930s and 1940s, a new shorter pulse time domain opened up as a result of the "bunched" nature of electron beams in cyclotron and synchrotron electron accelerators, due to the time and phase restrictions of repetitive geometry accelerating structures. Electron bunches of the order of picoseconds are now routine in the current generation of electron accelerators. However, these machines are usually very large and expensive (&gt;$10 million for low energy machines), and are currently confined to government laboratories for basic research applications.
The introduction of lasers (in the 1960s) has led to several techniques to produce x-ray pulses, based on the ability to focus laser beams to very high powers over small areas (microns). One recent technique involves short-pulse lasers to create a plasma by focusing the light onto a thin foil, which is heated to several thousand or more degrees over a region of several microns in a time very short compared to the thermal diffusivity of the foil. This plasma then radiates as a black body, producing an x-ray spectrum with energy spread dependent on the temperature of the plasma. However, this technique results in x-ray emission over 4 pi steradians, or isotropic radiation, and thus the x-ray flux drops rapidly with distance from the source (roughly as 1/r.sup.2). More importantly, the x-ray pulse length depends on the plasma thermodynamic properties, which are not easily controlled, and this results in x-ray pulses significantly longer than that of the laser driver, as demonstrated by the Umstadter, et. al. patent.
The use of laser accelerator electrons combined with novel x-ray conversion in this invention should provide advantages in the intensities of the x-rays produced and the angular distribution of x-rays. These features, coupled with the very short pulse duration will open up new capabilities for research and materials processing. For example, reaction kinetics and related molecular structure changes for a variety of important biological molecules could be studied, since many of these processes occur on picosecond time scales. Photosynthesis reactions in particular involve electron transfers at molecular sites that are known to be sub-picosecond processes. The very short x-ray pulses also can "freeze" molecular structures for x-ray diffraction studies. Another potential application is the study of the disordering and re-ordering (annealing) of semiconductor surface layers after rapid "melting" by short laser pulses followed by x-ray probes of the layer behavior. The effects of "hot electrons" injection in semiconductor devices could also be studied using the above laser techniques. Finally, changes in the above-K-edge absorption of x-rays (EXAFS) can be used to examine changes in short range structure at molecular sites, even for "amorphous" collections of molecules (as is the case in fluids, as opposed to crystalline structures).